YoMo

license

Apache-2.0(declared in Cargo.toml+README)

source

public

core-gated

ungated(the whole runtime is in src/ - router zipper serverless host tool invoker LLM and model providers auth and TLS - with no second license and no license key. Vivgrid sells a separate hosted serverless platform that the repo neither references nor requires)

Cargo.toml declares license = "Apache-2.0" at v2.0.6 and the README's License section points at the Apache License 2.0, but no LICENSE file is committed at the repository root, so the GitHub API reports the license as null. That is the weakest part of this record. Nothing is gated, though: what YoMo publishes is the whole runtime rather than a client to a hosted one. src/ carries router.rs, zipper.rs, serverless.rs, agent_loop.rs, tool_invoker.rs, tool_mgr.rs, llm_provider/, model_api_provider/, auth.rs, http_auth.rs, tls.rs and connector.rs - routing, the serverless host, tool invocation, the OpenAI-compatible surface, auth and transport - with no ee/ directory, no second license and no license-key path. Vivgrid sells "serverless LLM function calling that lets enterprise AI agents run their tools in the cloud - with observability, evaluation, and globally distributed inference built in", on a free tier plus credits, but its landing page does not mention YoMo and the current YoMo README does not mention Vivgrid. A hosted platform sold beside a complete open core is not a gate.

• https://github.com/yomorun/yomo/blob/main/Cargo.toml recorded 2026-07-09

  license = "Apache-2.0", name = "yomo", version = "2.0.4"

• https://github.com/yomorun/yomo recorded 2026-07-09

  README License section links Apache License 2.0; Rust core, QUIC, serverless SDKs public

• https://raw.githubusercontent.com/yomorun/yomo/main/Cargo.toml recorded 2026-08-12

  [package] name = "yomo", version = "2.0.6", edition = "2024", license = "Apache-2.0", description "A QUIC-based runtime for AI-LLM tool routing and serverless execution". The license is declared here rather than in a root LICENSE file, which is why the GitHub API reports license null for this repo.

• https://api.github.com/repos/yomorun/yomo/contents/src recorded 2026-08-12

  The published runtime, module by module: agent_loop.rs, auth.rs, bridge.rs, client.rs, connector.rs, http_auth.rs, llm_api.rs, llm_provider/, llm_stream_mapper.rs, metadata_mgr.rs, model_api.rs, model_api_provider/, openai_http_mapping.rs, router.rs, serve_config.rs, serverless.rs, tls.rs, tool_api.rs, tool_invoker.rs, tool_mgr.rs, zipper.rs and bin/. Routing, the serverless host, tool invocation, the OpenAI-compatible surface, auth and transport are all here; nothing is stubbed out to a hosted service.

• https://vivgrid.com/ recorded 2026-08-12

  Vivgrid sells "serverless LLM function calling that lets enterprise AI agents run their tools in the cloud - with observability, evaluation, and globally distributed inference built in", with a free tier and $200/month in credits for maintainers of approved open-source projects. The landing page does not mention YoMo at all, so what is sold is a hosted platform beside the framework rather than a piece of it.

1,920 GitHub stars on yomorun/yomo, which bands at 1K-10K stars, level 2 on the stars scale - a scale that caps at 3 in any case, because a star is not a use. No download, install or customer figure is published for this product: crates.io shows roughly 250 downloads a month, and there is no mass-market usage volume to band. Stars are the only honest signal available here, so the level is directional.

• https://api.github.com/repos/yomorun/yomo recorded 2026-08-12

  stargazers_count = 1,920 for yomorun/yomo

Distinctive full agent runtime with an edge/geo-distributed angle and multi-language SDKs. Functional and coherent feature set for the agent-tooling layer but narrower ecosystem and integration surface than the category leaders (MCP itself, FastMCP). Rated 3 on feature breadth from repo/README review.

• https://github.com/yomorun/yomo recorded 2026-08-13

  Repo page - "Serverless AI Agent Framework with Geo-distributed Edge AI Infra", section "Focuses on Geo-distributed AI Inference Infra", worked example posting to a serverless function at /tool/get-weather. Declared topics: a2a-protocol, agent-framework, distributed-cloud, function-calling, geodistributedsystems, managed-skills, mcp, openai, quic, realtime, serverless, skills. 1,920 stargazers.